New Evidences on the Disposition Effect of Individual Investors

New Evidences on the Disposition Effect of Individual Investors Cristiana Cerqueira Leal Lecturer in Finance Management Department Minho University - Portugal [email protected] Phone: +351-253-60-4561 Fax: + 351-253-28-4729 Manuel J. Rocha Armada Professor of Finance Management Department Minho University - Portugal [email protected] Phone: + 351-253-60-4555 Fax: + 351-253-28-4729 João L. C. Duque Professor of Finance ISEG – Lisbon - Portugal [email protected] Phone: + 351-21-392-5955 Fax: + 351-21-392-2808 This version: March 2006 New Evidences on the Disposition Effect of Individual Investors Abstract Financial theory has identified the tendency of investors to hold loosing investments too long and sell winning ones too soon. This tendency was denominated the disposition effect by Shefrin and Statman (1985). This research provides evidence of the disposition effect on the Portuguese stock market, by studying a unique database that consists on trading records of 1496 individual investors. The preference for realising gains to losses was observed every month of the year and for all individual investors. Even at the end of the fiscal year, the disposition effect still holds (in spite of the existence of fiscal incentives for the so-called fiscal effect), as opposed to the evidence found in other markets. We also found that more sophisticated investors (classified on the basis of frequency of transaction, volume and portfolio value), are less prone to the disposition effect than less sophisticated ones. We also identified differences as to the disposition effect in terms of its intensity, when considering different market trends. In bull market periods, the disposition effect is even more evident than in bear markets and significant in both markets. 2 New Evidences on the Disposition Effect of Individual Investors 1. Introduction Financial theory has identified the tendency of investors to hold loosing investments too long and sell winning ones too soon. This tendency was denominated the disposition effect by Shefrin and Statman (1985). According to several financial theories, buying and selling decisions should be taken based on price expectations. Then, comparing historical prices, namely acquisition prices, with current prices is not a rational criteria for deciding to hold or to sell. If market efficiency holds, even in its weak form, past prices should not be relevant to resource allocation decisions. The main aim of this study is to analyse the disposition effect of individual investors using the Portuguese stock market and analysing a unique database of 1496 Portuguese individual investors’ trades, from 1st January 1999 to 31st December 2002 (159 406 trades). We can state four central questions to which we intend to answer in this study: - Are Portuguese individual investors prone to the disposition effect? - Does the fiscal effect reduce or invert the effect? - Are there significant differences in the disposition effect motivated by market tendency, i. e., is the disposition effect level different for bull and bear markets? - Which investors are more prone to the disposition effect, that is, does investors’ sophistication affect the disposition effect? We found the disposition effect for the entire period of the study, even in the end of fiscal year, suggesting that fiscal effect seems to have no significant impact in investors’ preferences. We also found that in bull markets the disposition effect is stronger than in bear markets and that more sophisticated investors are less prone to the disposition effect. This paper brings new evidence about disposition effect. Firstly, as we know, Portuguese market is the first studied market where fiscal effect does not reduce disposition effect by the 3 end of fiscal year, despite the incentives to do so. The legislation is not homogeneous for the investors within the sample and there were changes in fiscal law during the sampling period. Even tough, the principal explanation is that mental account plays a role that blinds investors for the real reasons for realising losses: they want to close up the year with good performance and do not take taxes into consideration. Then, we relate disposition effect with market mood. Kim and Nofsinger (2002) analyse the behaviour and performance of individual investors in Japan and found that trading behaviour varies depending on bull or bear market conditions. Based on this evidence we aim to analyse if also disposition effect depends on market mood. The disposition effect found in the Portuguese stock market is stronger than in other studied markets. This raises the question in which way Portuguese investors are different form other markets in a way that explains these difference. Several studies that take into consideration several classes/groups of investors give indications that investors do not exhibit disposition effect in the same degree. Grinblatt and Keloharju (2000) show that the selling behaviour is associated with the investor’s sophistication level and with their investment size. The most sophisticated investors have larger investments and pursue momentum strategies which results in superior performance. Moreover, Odean (1998) found differences in disposition effect for frequent and infrequent traders. Based on this evidence, we also intend to analyse if investors’ sophistication has impact in the disposition effect. This paper is organised as follows: first we introduce the nature of the disposition effect as well as some literature on previous studies. Then we present the database that was used in this study. It follows a presentation of the methodology and a discussion of the empirical results. We conclude with a summary of the paper. 2. The Disposition Effect and Previous Studies Disposition effect concerns to patterns of gains and losses realization. The explanations to the preference for realising winners than losers are generally based on behavioural factors emerging from Kahneman and Tversky’s (1979) “Prospect Theory” and Thaler’s (1985) “Mental Accounting Theory”. 4 Prospect theory describes how people evaluate their wealth based in psychological aspects. Within this framework, the most important to human perception is the difference of stimulus rather than their absolute magnitude, as alleged by Expected Utility Theory. This means that perception is a relative processus rather than an absolute one. According to the Prospect Theory individuals codify their wealth changes in terms of gains and losses using a reference point. Results above the reference point are seen as gains while results below are seen as losses. Usually, the reference point is the current status of wealth and a change in the reference point has an impact on the classification of gains or losses. It can also affect options formulation and decision making. According to the Prospect Theory, individuals also exhibit decreasing sensitivity to outcomes. This means that when gains or losses are distant from the reference point they lose significance. Once investors usually take the acquisition price as its reference point, accounting gains and losses anchored on it, the first euro gained will be recognised as being more valuable than the second, the second one as being more valuable than the third one, and so on. Identically, the pain associated with the loss of first euro will be recognised as being stronger than the loss of the second one, the second stronger than the third, and so on. As a result of this approach, marginal stock price increases of the same magnitude will provide decreasing marginal recognised value. On the contrary, when stock prices fall below the reference point, the first loss represents a greater pain than the subsequent ones. This non-linear decreasing function will tend asymptotically to a flat region of the recognised value function. However, this recognised value function changes in shape if prices restart raising after an initial fall. When prices restart raising a new reference point is established in the investor’s mind. A new and reshaped recognised value function branch, with a positive but declining marginal recognised value starts. Therefore, while additional price decreases will still be treated through the original value function being and considered less penalising because they are placed in a flatter region of the value function, the same price decrease following a significant raise results into a very different story. With the change of the reference point to a higher value we would observe a shift on the recognised value function. The decrease after the raise would now coincide with a high sensitivity area of the recognised value function Investors would than exhibit an 5 increasing sensitivity to recovery with a stepper zone of value function. This explains why investors prefer to realize gains and to defer losses. They prefer to realize a gain because marginal gains are recognised as less valuable that possible marginal losses of the same amount. On the contrary, when they are in the losses zone, additional losses will not be recognised so painfully, while a possible recovery has a greater value. Investors are not very sensitive to additional losses but are very sensitive to possible price recoveries. Another fundamental aspect of disposition effect analysis relates to the investors’ nonportfolio approach. Thanks to the mental accounting process individuals segregate wealth in numerous accounts Investors tend to keep them separated by mental barriers and exhibit different behaviour for each account. The disposition effect, based on mental accounting, considers each stock individually treated instead of considering them a part of a portfolio. When a new stock is bought, a new mental account is opened with its reference point, from which gains and losses are calculated and disinvestment decisions are made. While stocks are hold gains or losses are not considered real. The disposition effect has been found for many markets and data periods. Based on US investors’ trading records, Schlarbaum, Leweleen and Lease (1978) found evidence for retail brokerage clients from 1964 to 1970; Odean (1998) found evidence to discount brokerage clients from 1987 to 1993 and Locke and Mann (1999) document the same behaviour to professional futures traders in 1995. The disposition effect was also identified for home buyers and sellers by Case and Shiller (1998) and Genesove and Mayer (2001). This preference was also found in other markets, namely, by Brown, Chappel, Rosa and Walter (2002) in Australian Stock Exchange, and Shapira and Venezia (2001) among professional investors in Israel. Grinblatt and Keloharju (2000) found that Finland’s domestic investors pursue contrarian behavior and Kim and Nofsinger (2002) found a preference for selling past winners in Japan, which is consistent with being disposition-prone. In terms of measurement methods we can identify three approaches based on the data format used: market data; portfolio data and experimental markets. 6 The methods based in market data (market perspective) compare volume and market prices’ changes (e.g.: Dyl (1977); Lakonishok and Smith (1986); Ferris, Haugen and Makhija (1988) and Kaustia, 2000). In general, the purpose is to identify whether volume changes are motivated by winners or losers. If disposition effect holds, it is expected higher volumes for bulish than for bearish times, which means the existence of a preference of winners to sell. The methods based in portfolio data (investor perspective) allow a deeper and accurate analysis. It turns possible to look in detail into each investor portfolio and check whether the stocks sold are the winning or the loosing ones (e.g.: Schlarbaum, Leweleen and Lease (1978); Odean (1998) and Brown, Chappel, Rosa and Walter, 2002). Finally, the methods based on experimental markets attempt to reproduce the stock trading to assess the preference for holding loosing investments while selling winning ones (see Weber and Camerer (1998); Chui (2001) and Oehler, Heilmann, Lager and Oberlander, 2002). 3. Data This study is based on a unique database of 1496 individual investors’ accounts with detailed data on their registered trades. The data set goes from 1st January 1999 to 31st December 2002 and comprises 159 406 trades. In order to ensure that the accounts represent the entire stock portfolio for each investor, we only consider investors that trade exclusively in Portuguese market. Otherwise we would, eventually, be considering partial accounts. According to CMVM (2003) the national market represents the main destiny of the security investment for Portuguese investors (94.3%). The data was provided by a discount brokerage firm with a market share of about 1% of Portuguese market in the sampling time period. The analysis considers 1496 investor accounts that trade at least once in the sampling time period. The data is composed by initial value positions, initial quantity positions, account movements both in value, and in volume, events, and daily closing stock prices. The files include all movements that occurred in each account within the time period. We have excluded non security data, namely, on bonds and on warrants. We excluded accounts of investors that trade derivatives, since these could act on the underlying asset for hedging or arbitrage purposes that would undermine our analysis. The 7 most important files for the analysis are account movements in value and account movements in quantity. The database includes 159 406 trades (81 914 executed buys and 77 492 executed sells). This means an average of 106.6 trades per account for the entire period and an average of 159.4 trades per trading day for the entire set of accounts. Based on this information, we have reconstructed each investor account for each day of the sampling period. We have net all trades on the same day and asset for the same investor and ignored all sells to which it was not possible to identify the purchase date and its price (purchases before 1st January 1999, because lack of information). We have also corrected the data for stock splits, mergers and acquisitions. 4. Methodology We start by a simple test checking whether Portuguese individual investors exhibit the disposition effect (as defined by Odean (1998)). If this is true, we expect to find empirical evidence that the Proportion of Gains Realized is superior to the Proportion of Losses Realized in the sampling period. In order to accept this hypothesis, we will test the null hypothesis under which the proportion of investors’ realised gains isn’t greater than realized losses, that is: H0: Proportion of Gains Realized ≤ Proportion of Losses Realized. H1: Proportion of Gains Realized > Proportion of Losses Realized. Then we analyse if the disposition effect is affected by the fiscal effect. If the fiscal effect holds in the Portuguese market, in December investors would exhibit a preference for realizing more losses than gains and this would be rather distinct from the other months of the year calendar. However, and contrary to the evidence in other markets, if hedonic reasons hold we would expect opposite evidence. If investors tend to realise gains to close up the year with a good performance then, the fiscal effect if noticed, would not have enough significance in order to reduce or invert the disposition effect. Consequently, we will test the hypothesis that the pattern of gains and losses realization on December is not significantly different from the other months of the year. 8 H0: Proportion of Losses Realized in December - Proportion of Gains Realized in December > Proportion of Losses Realized from January to November - Proportion of Gains Realized from January to November. H1: Proportion of Losses Realized in December - Proportion of Gains Realized in December ≤ Proportion of Losses Realized from January to November - Proportion of Gains Realized from January to November. These two tests for global disposition effect and for the difference on disposition effect for data partitions were defined by Odean (1998) and are generally found in the literature that studies disposition effect based on investors’ portfolio thereafter. On this basis, we will also analyse the disposition effect related to the market mood and the disposition effect related to investor’s sophistication. In the literature Kim and Nofsinger (2002) based on market data analyse the behaviour and performance of individual investors in Japan and found that trading behaviour varies depending on bull or bear market conditions. Then, it is relevant to study whether disposition effect is affected by market mood. Once our analysis is based on investor portfolio data the behaviour measure will be much more reliable. Then, the third question that we intent to address relates the disposition effect with market mood. There are more opportunities of gains realisation during bull periods. Therefore, we should notice this fact in our database, if disposition effect is dependent upon the market mood. The measure of the disposition effect we used, considers the realization of gains and losses compared to the opportunities that exists due to market tendency. Consequently, the methodology is suitable to measure differences in bull and bear periods. We will assume bull periods those where daily market capitalization is consistently going up and bear periods those where the daily market capitalization is consistently going down. Then, the bull market period is from 1st January 1999 to 3rd March 2000 and the bear market period is from 4th March 2000 to 31st December 2002. Then the third hypothesis compares the disposition effect in each period: H0: Proportion of Losses Realized in bull periods - Proportion of Gains Realized in bull periods > Proportion of Losses Realized bear periods - Proportion of Gains Realized on bear periods. 9 H1: Proportion of Losses Realized in bull periods - Proportion of Gains Realized in bull periods ≤ Proportion of Losses Realized in bear periods - Proportion of Gains Realized in bear periods One could expect that in bull periods the disposition effect should be attenuated or even inverted if investors follow momentum strategies. Nevertheless, it is harder to realise losses during bull periods as a result of behavioural reasons. When prices tend generally to raise, losses realisation affects investors’ confidence and in order to hide mistakes investors tend to keep loser investments hoping their recovery. Finally, we will also analyse whether individual investors are equally prone to the disposition effect whatever the level of sophistication they show is. We try to identify the sophistication by segregating investors into different groups according to the number of trades they execute, the trading volume that their trades involve and according to the portfolio value (wealth). Some studies give indications that different groups of investors have different degrees of disposition effect. Grinblatt and Keloharju (2000) show that the most sophisticated investors have larger investments and pursue momentum strategies while less sophisticated ones follow contrarian strategies. This suggests that less sophisticated investors should denote stronger disposition effect. Dhar and Zu (2002) found that wealthier investors and investors with trading experience exhibit less disposition effect. Furthermore, Odean (1998) found intensity differences in disposition effect for frequent and infrequent traders even though booth groups show preference for sell winning investments. Based on this, we will analyse the impact of investors’ sophistication in the disposition effect. Finally, Brown, Chappel, Rosa and Walter (2002) found that more sophisticated investors (considering value of transaction as sophistication criteria) express lower disposition effect, although even large traders prefer hold their losing investments and sell their winning ones. The segregation by the number of trades allows the comparison of investors based on their trading frequency. We assume that more active investors tend to be more sophisticated. However, this figure ignores the trading volume per trade, which means that trading one share would equal the significance of trading one thousand. In order to take theses differences into account we will also test differences in behaviour based on the number of shares traded, assuming that more sophisticated traders tend to present higher trading volume per trade. 10 Finally we also use the average portfolio value within the sampling period which may be, probably, the best criterion to identify investors’ sophistication. For each criterion, we will test if there are significant differences in the disposition effect by dividing investors into two groups. This means the following hypothesis: A) Using the number of trades: H0: Proportion of Losses Realized by low frequency traders - Proportion of Gains Realized by low frequency traders ≤ Proportion of Losses Realized by high frequency traders - Proportion of Gains Realized by high frequency traders. H1: Proportion of Losses Realized by low frequency traders - Proportion of Gains Realized by low frequency traders > Proportion of Losses Realized by high frequency traders - Proportion of Gains Realized by high frequency traders. B) Using the trading volume per trade: H0: Proportion of Losses Realized by low volume traders - Proportion of Gains Realized by low volume traders ≤ Proportion of Losses Realized by high volume traders - Proportion of Gains Realized by high volume traders. H1: Proportion of Losses Realized by low volume traders - Proportion of Gains Realized by low volume traders > Proportion of Losses Realized by high volume traders - Proportion of Gains Realized by high volume traders. C) Using the trading volume per trade: H0: Proportion of Losses Realized by low portfolio value - Proportion of Gains Realized by low portfolio value ≤ Proportion of Losses Realized by high portfolio value - Proportion of Gains Realized by high portfolio value. H1: Proportion of Losses Realized by low portfolio value - Proportion of Gains Realized by low portfolio value > Proportion of Losses Realized by high portfolio value - Proportion of Gains Realized by high portfolio value. As mentioned by Odean (1998), in order to test if the investors have the tendency for selling winners too soon while holding losers too long, we have to take into consideration the impact of market trends. In a bullish market, where the majority of share prices raise, , there are more winning stock. Consequently, we observe more opportunities to sell winners than losers, even though investors may be indifferent in selling winners or losers. On the contrary, in a bearish 11 market, with a large number o share prices falling, investors have more loser stocks in their portfolios and, as a result, in the assumption that they are indifferent in selling winners or losers, it is expected they sell more losers. Again, this may have nothing to do with their human nature but is just the result of the market downward, creating more opportunities to sell. Therefore, it is necessary to remove the impact of market tendencies. In order to detect investors’ behavioural tendencies without the market effect, the disposition effect will be identified and measured taking into account the selling of winners and losers relative to the potential opportunities for selling winners and losers that are hold in their accounts. We start by computing Realised Gains and Realised Losses. Realised Gains (RG) and Realised Losses (RL) in a security i, in a specific account l, are calculated as the difference between the selling price and the reference price (the average acquisition price for that security). RGi, l = S i, l − Ai, l > 0 RLi, l = S i, l − Ai, l < 0 where S i , l represents the selling price for security i in account l, and Ai , l represents the average acquisition price of security i in account l. The average security price is a weighted average considering the number of shares bought in each buying transaction. If an account registered k buying trades on security i before being sold, with trades composed by n shares bought at price P, the average security price become: k ∑ n j Pj Ai, l = j =1 k ∑nj j =1 If only one buying trade occurred, the average security price is the sole buying price registered in that account for that specific security. Realised Gains and Losses are only computed when a selling trade occurs for a security in an account. Potential Gains (PG) and Potential Losses (PL) in an account are only calculated when a selling trade is registered in that account. That is, we only care about Potential Gains or Losses when a selling trade occurs. Potential Gains and Potential Losses are calculated as the difference between the reference price (the average acquisition price) and the closing price of the day when security i is sold: 12 PGi, l = Ci, l − Ai, l > 0 PLi, l = Ci, l − Ai, l < 0 where Ci , l represents the closing price for security i in account l on the day when security i was sold. As we did for Realised Gains and Losses, Potential Gains and Potential Losses are only computed when a selling trade occurs for a security in an account. We calculated Realized Gains, Realized Losses, Potential Gains and Potential Losses for each day there is one or more sells in an account that has at least two securities and that does not sell the entire portfolio in that day. Otherwise, we can not calculate Potential Gains and Losses as there are no residual securities that could potentially be sold. Then, we calculated the Proportion of Gains Realized (PGR) and the Proportion of Losses Realized (PLR) as follows: N ∑ RG i i =1 PGR = l, t M N ∑ RGi + ∑ PG j j =1 i =1 N ∑ RLi i =1 PLR = l, t N M ∑ RLi + ∑ PL j i =1 j =1 where PGRl ,t represents the Proportion of Gains Realised in account l on day t, RGi stand for the Realised Gain on security i on day t. As we only calculate Realised Gains and Potential Gains when sells occur, N represents the number of securities sold in day t, and M represents the residual number of securities kept intact in account l. The same methodology was used for PLRl ,t that stands for Proportion of Losses Realised. As in Odean (1998), we use t-test for testing the statistical significance of the differences in the proportions PGR and PLR. A significant negative difference means that investors exhibit a preference to hold losing investments and to sell the winning ones. In other words, if a significant and negative difference is found, the disposition effect exists. 13 The standard error for the difference in the proportions PGR and PLR is given by: σ (PRL − PGR ) = PGR(1 − PGR) PLR( 1 − PLR) + N RG + M PG N RL + M PL Where N RG , M PG , N RL and M PL stand, respectively, for the number of Realized Gains, Potential Gains, Realized Losses and Potential Losses. In order to segment investors, we will divide the database in percentiles using the following criteria: trading frequency; trading volume and average portfolio value in the sampling period. The next question is which point should be used as the division point. We divided investors into two different groups testing differences for the following percentiles: 50%, 75%, 90% and 95%. 5. Discussion of Empirical Results 5.1. The Disposition Effect and the Fiscal Effect We find evidences of strong disposition effect for the entire sample with a difference between the Proportion of Gains Realized (PGR) and the Proportion of Losses Realized (PLR) of about 0.21. Table 1 shows the results for PGR and PLR first for the entire year, and second segregating results from January to November from those observed in December. Those tests are based on the number of realized gains, number of realized losses, number of potential gains and number of potential losses. These observations are aggregated for all investors and throughout time, assuming that these observations are independent across investors and time. 14 Table 1 – Global Disposition Effect This table reports the Proportion of Gains Realized (PGR) and the Proportion of Losses Realized (PLR) from 1-1-1999 to 31-12-2002. The number of Realized Gains (RG), Realized Losses (RL), Potential Gains (PG) and Potential Losses (PL) are aggregated across accounts and days. T-statistics assume that GR, LR, PG and PL result from independent decisions. RG RL PG PL PLR PGR PLR-PGR PGR/PLR σ(PLR-PGR) t-statistic Entire Year 25 891 21 629 55 553 176 975 December 2 334 1 363 5 597 12 829 Jan.-Nov. 23 557 20 266 49 956 164 146 0.109 0.318 -0.209 2.919 0.00178 -117.735 0.096 0.294 -0.198 3.064 0.00568 -34.881 0.110 0.320 -0.211 2.916 0.00187 -112.663 The differences for PGR and PLR are significant for the entire year. However, contrarily to the empirical evidence from similar studies (such as Odean (1998), Brown, Chappel, Rosa and Walter, 2002), we found the disposition effect even in the end of fiscal year. The difference between PGR and PLR in December is 0.2, only slightly lower than the rest of the year. When testing if the difference between PGR and PLR in December is significantly different from that obtained from January throughout November (i. e., PLR-PGR in December ≤ PLR- PGR in January - November), we get a t-statistic of 2.165. Therefore, we conclude that this difference is not significant for a confidence level of 0.011. Moreover, the preference to sell winners in December is three times superior to the preference to sell losers. This is even higher than during the period from January to December. Then, we can conclude that in the period, investors show disposition effect for the entire year, including December. The difference found between PGR and PLR is clearly higher than the evidence of other similar studies (such us Odean, 1998). We believe that this is due to the relatively low level of sophistication of the individual investors. Even though the Portuguese stock market stoped to be considered an emergent market, it still exhibits some characteristics of those markets, namely higher volatility and longer time for crisis recovery. Investors in such markets tend to reveal low sophisticated behaviour, particularly individual investors. Grinblatt and Keloharju (2000) state that Finish investors, specially, individual ones behave with low sophistication, being prone to contrary behaviour and consequently to the disposition effect. 1 Given the possibility of the test being inflated due to the lack of independence of the parameters, we demand a high confidence level. 15 As opposed to the evidence found in other markets, we observed the disposition effect for the entire year, even in December (the end of fiscal year), in spite of the existence of fiscal incentives to engage in tax-motivated selling. It was already expected that fiscal effect had a low impact in realised gains and losses first because the legislation is not homogeneous for the investors within the sample, and second because there were changes in fiscal law during the sampling period. Even if some investors decide to realize capital losses for tax purposes, there are other investors which are not motivated by fiscal aspect and will prefer realize capital gains at the end of the year to evidence improved performance. Obviously, these acts result from mental accounting mechanisms that only consider gains those already realized. Therefore, the fiscal effect and the disposition effect counterbalance themselves and the pattern of gains and losses realization is not significantly modified in December, for the Portuguese market. Another possible explanation is that investors do no take taxes into consideration at all and consequently the pattern of gains and losses holds in December. As can be observed in graph 1 there are no change in PGR/PLR in December. Notwithstanding the taxation of capital gains, it is estimated that a larger amount of individual investors do not declare capital gains. Even those who declare capital gains for taxation purposes are blinded by mental accounting and do not consider this variable in gains and losses computation. Although we found evidence of the disposition effect for the Portuguese individual investors, it is relevant to investigate if it is constant during the sampling period. Table 2 reports the PGR and PLR for each year in analysis. We found the same disposition effect for every year, as we found for the entire period. Table 2 - Disposition Effect for Each Year This table reports the Proportion of Gains Realized (PGR) and the Proportion of Losses Realized (PLR) for each year over the period from 1-1-1999 to 31-12-2002.The number of Realized Gains (RG), Realized Losses (RL), Potential Gains (PG) and Potential Losses (PL) are aggregated across accounts and days. T-statistics assume that GR, LR, PG and PL result from independent decisions. RG RL PG PL PLR PGR PLR-PGR PGR/PLR σ(PLR-PGR) t-statistic 1999 8 893 5 095 17 809 46 861 2000 11 354 8 334 24 508 70 442 2001 3 231 4 899 7556 35434 2002 2 413 3 301 5680 24238 0.098 0.333 -0.235 3.396 0.00317 -74.230 0.106 0.317 -0.211 2.993 0.00269 -78.378 0.121 0.300 -0.178 2.466 0.00470 -37.881 0.120 0.298 -0.178 2.487 0.00545 -32.722 16 Graph 1 details the evolution of the ratio PGR to PLR. The preference to realize winners is at least 2.5 times superior to the preference to realize losers for the entire sampling period. The ratio varies between 2.5 and 3.3. Once again, we observe how fiscal effect has little influence into the pattern of the realization of gains and losses, showing a growing preference for realising winners at the year end. February, July and August are the months that evidence the highest ratio while March, April, September and October have the lowest. In terms of the difference between PGR and PLR, for every month the values are closer to 0.2, corroborating that, consistently, investors prefer realizing winners than losers. Graph 1 – Evolution of PGR/PLR Aggregated by Month December November October September August July June May April March February 3,50 3,00 2,50 2,00 1,50 1,00 0,50 0,00 January PGR/PLR This graph reports the ratio Proportion of Gains Realized (PGR)/Proportion of Losses Realized (PLR) aggregated by month from 1-1-1999 to 31-12-2002. The number of Realized Gains (RG), Realized Losses (RL), Potential Gains (PG) and Potential Losses (PL) are aggregated across accounts and days. These tests were developed under the assumption that the number of realized gains, the number of realized losses, the number of potential gains and the number of potential losses are independent across investors and time. Although this assumption does not bias the tests, it can inflate them. As the t-statistics we get are very high this is not problematic to the analysis. In spite of this, in order remove any doubt about the effect, we will, additionally look to an alternative test. Firstly, we started by assuming that independence only exists throughout time. In order to overcome the problem with investors’ dependence, we calculate PGR, PLR and its difference per investor. Then, we calculate the average PGR, and the average PLR and PLR-PGR and we test the average difference. The results for this test are shown in table 3: the PGR is 0,57; the PLR is 0.21 and the difference 0.36. By comparison to the previous test (table 1), we 17 conclude that in this alternative test both the proportions and the differences found are higher, and so is the t-statistic. As a result, in this alternative test the null hypothesis was also rejected with a t-test statistic of 184. The same happened with the December data, that is, the average PGR is significantly superior to PLR, exhibiting a t-statistical even higher than that in previous test. Table 3 – Disposition Effect – Alternative Measure to Control Accounts Dependence This table reports the Proportion of Gains Realized (PGR) and the Proportion of Losses Realized (PLR) from 1-1-1999 to 31-12-2002. The number of Realized Gains (RG), Realized Losses (RL), Potential Gains (PG) and Potential Losses (PL) are aggregated across days. The tstatistics calculation assumes that GR, LR, PG and PL result from independent decisions over time. RG RL PG PL PLR PGR PLR-PGR PGR/PLR σ(PLR) σ(PGR) σ(PLR-PGR) t-estatistic Entire Year 29.03 24.19 62.28 197.96 December 6.05 3.35 14.50 31.52 0.205 0.566 -0.360 2.755 0.17379 0.24148 0.00196 -183.985 0.167 0.499 -0.333 2.994 0.18999 0.31106 0.00643 -51.740 Surprisingly, the test to control the lack of investors’ independence reveals a stronger statistical significance. The higher results for PGR, PLR and for the differences and, consequently, the higher t-statistic, can be explained by the equal weight assigned to each investor in this alternative measure. Each investor has the same weight in the calculation of the disposition effect, independently of the high or low level of transaction in its account. These results suggest that investors, which trade less frequently, exhibit stronger disposition effect. When we weight them more heavily, PGR, PLR and its difference becomes higher. This means that PGR and PLR are dependent of the investors to which they are calculated. This fact can be considered critical to the alternative test, once that the accounts with high frequency transaction provide more accurate estimative for the calculation of the proportions. Nevertheless, this test satisfies its purpose of controlling if the statistic test is inflated and shows that we can accept the conclusion provided by the initial test. We also need to control the impact of decision dependence over time. For that purpose, we will ignore consecutive sells. Whenever we find more than one sells within a week (five 18 consecutive trading days) on the same stock, only the first sell will be considered. In a similar way, we only consider the first sell within the week to calculate potential gains and losses. The test is done, once again, by calculating the PGR and the PLR for each investor and then the average for all investors. Table 4 reports the results. We conclude that after controlling for time dependence, the disposition effect is still evident and seems even stronger. The difference between PGR and PLR is slightly higher than that observed in the previous test. This procedure does not guarantee that decisions are independent over time. When investors hold loser stocks the decision can be kept for a period longer than 5 days. Especially in longer bear periods, one can expect that investors hold their loser investments for longer periods, expecting to recover their losses in order to achieve the break-even. As we saw previously, the satisfaction provided for this possibility is higher than the pain imposed by additional losses. This is one of the reasons why investors accept risky bets. And in a difficult situation, it is always preferable to hold the status quo, once a non-decision is not as painful as a wrong decision. Table 4 – Disposition Effect – Alternative Measure to Control Accounts and Time Dependence This table reports the Proportion of Gains Realized (PGR) and the Proportion of Losses Realized (PLR) over the period from 1-1-1999 to 3112-2002. The number of Realized Gains (RG), Realized Losses (RL), Potential Gains (PG) and Potential Losses (PL) are aggregated across days for each investor. RG RL PG PL PLR PGR PLR-PGR PGR/PLR σ(PLR) σ(PGR) σ(PLR-PGR) t-estatistic Entire Year 11.07 8.50 12.69 34.66 December 3.39 1.90 4.43 9.08 0.263 0.660 -0.397 2.510 0.22092 0.26941 0.00527 -75.369 0.247 0.583 -0.336 2.363 0.30918 0.32713 0.01872 -17.953 Up to the moment PGR and PLR are calculated on the basis of the number of sells and potential sells. Even though, to measure the impact of investors’ dimension is pertinent to calculate these proportions on the basis of volume and value of the traded stock. 19 When the number of sells and potential sells is taken as basis, all transactions are considered equally important and they are used with the same weight when calculating our instrumental variables. However, it is pertinent to ask whether the disposition effect is still persistent when we take into consideration the number of stocks sold and the number of stocks that could have been sold. Table 5 reports the results for PGR and PLR calculated on the basis of stocks instead of singular sells. The proportions as well as its difference are now lower. Nevertheless, the difference is still significant (t-statistic of 89). The lower PGR and PLR can be explained in line with the explanations presented for the alternative tests. Investors with higher transaction volume (and more frequent trades) are now strongly weighted and these investors exhibit lower PGR and lower PLR as well as a lower difference between these proportions. This means they are less prone to the disposition effect. Table 5 – Disposition Effect Based on Volume This table reports the Proportion of Gains Realized (PGR) and the Proportion of Losses Realized (PLR) over the period from 1-1-1999 to 3112-2002. The volume of Realized Gains (RG), Realized Losses (RL), Potential Gains (PG) and Potential Losses (PL) are aggregated across accounts and days. The t-statistics calculation assumes that GR, LR, PG and PL result from independent decisions. RG RL PG PL PLR PGR PLR-PGR PGR/PLR σ(PLR-PGR) t-statistic Entire Year 16 764 347 20 927 061 58 433 585 233 432 824 December 1 581 558 1 626 243 6 391 875 20 006 371 Jan.-Nov. 15 182 789 19 300 818 52 041 710 213 426 453 0.082 0.223 -0.141 2.710 0.00158 -88.835 0.075 0.198 -0.123 2.639 0.00499 -24.661 0.083 0.226 -0.143 2.723 0.00167 -85.551 Hypothetically, investors may realize gains and losses of significantly different amounts. Therefore, we also calculated the PGR and PLR based on the value of gains and losses. This is critical, because more import than a gain or loss should be its magnitude. If realized small gains are frequent but big losses are seldom, the previous conclusions could be challenged. Table 6 shows the results for value weighted PGR and PLR and the conclusions are similar to previous, based on trading volumes: the PLR is 0.06; the PGR is 0.18 and its difference has a t-statistical of 81.5. Considering the value of gains and losses we get lower figures for the proportions. This is likely because investors with low amounts in transactions are more prone to the disposition effect. Nevertheless, we keep concluding for the disposition effect for the total data set and for every month, including December. Brown, Chappel, Rosa and Walter 20 (2002), when considering values of transaction, also get the same conclusions for individual investors, although for other investors’ classes disposition effects disappears on this calculation. This indicates that, when pondering strongly the larger investments, investors with more experience/sophistication are more weighed and these are not so eager for gains realization (the reduction in disposition effect comes specially from the reduction of PGR, when compared with volume). Attending the monthly progress of the ratio PGR/PLR calculated on the basis of volume and value, we keep concluding that the preference for selling winners is always, at least, twice than the preference for selling losers. Table 6 – Disposition Effect Based on Value This table reports the Proportion of Gains Realized (PGR) and the Proportion of Losses Realized (PLR) over the period from 1-1-1999 to 3112-2002. The value of Realized Gains (RG), Realized Losses (RL), Potential Gains (PG) and Potential Losses (PL) are aggregated across accounts and days. The t-statistics calculation assumes that GR, LR, PG and PL result from independent decisions. RG RL PG PL PLR PGR PLR-PGR PGR/PLR t-statistic Entire Year December Jan.-Nov. 8 156 844 15 678 820 36 982 974 236 451 301 686 055 844 335 4 085 564 15 379 937 7 470 789 14 834 485 32 897 410 221 071 364 0.062 0.181 -0.119 2.906 -81.562 0.052 0.144 -0.092 2.763 -21.047 0.063 0.185 -0.122 2.943 -79.348 As a conclusion, we may state that individual investors in the Portuguese market exhibit the disposition effect for the entire year, including December. The preference to realize winners instead of losers is stronger in our study than in other studies, probably because Portuguese individual investors are less sophisticated than in other more developed markets. Opposed to other markets studied, the disposition effect is neither inverted nor reduced in the end of the fiscal year. 5.2. The disposition effect in bull and bear markets In bull periods, there are more opportunities for realising gains, while in bear periods there are more opportunities for realising losses. However, the measure of the disposition effect under use already considers this potential trap. Realised gains and losses are computed in 21 comparison to the existent opportunities due to the market trend. Therefore, if the measure used to compute the disposition effect is not affected by the market trend the differences that may arise if any are the result of investors’ preferences under different market conditions. We split the period of analysis into a bull market period (from 1st January 1999 to 3rd March 2000) a bear market period (from 4th March 2000 to 31st December 2002) and repeated the previous tests. Table 7 shows the PGR and PLR for each sub-period, calculated on the basis of the number of sells. We found strong evidence of the disposition effect both in bull and in bear markets, but stronger in bull markets. The difference found between the disposition effect in bull and bear markets is statistically significant. Table 7 – Disposition effect in bull and bear markets This table reports the Proportion of Gains Realized (PGR) and the Proportion of Losses Realized (PLR) over the bull period (from 1st January 1999 to 3rd March 2000), the bear period (from 4th March 2000 to 31st December 2002) and its difference. The number of Realized Gains (RG), Realized Losses (RL), Potential Gains (PG) and Potential Losses (PL) are aggregated across days. The t-statistics calculation assumes that GR, LR, PG and PL result from independent decisions over time. RG RL PG PL PLR PGR PLR-PGR PGR/PLR σ(PLR-PGR) t-statistic [(PLR-PGR)Bull – (PLR-PGR)Bear] t-statistic Bull 13 890 6 494 28 270 62 065 0.095 0.329 -0.235 3.478 0.0025 -92.138 -0.046 -17.909 Bear 12 001 15 135 27 283 114 910 0.116 0.305 -0.189 2.625 0.0025 -76.000 For illustration purposes and deeper detail when studying the differences between bull and bear markets, the graph 2 exhibits the monthly progress of the PGR/PLR ratio and the graph 3 shows the PGR, PLR and its difference for each month. Considering the graph 2, it shows clearly that, in the bull period, the preference for realising gains is, at least, three times higher than the preference for realising losses, while in the bear period, the ratio falls (even though, the tendency for realizing gains is still at least twice than the tendency for realizing losses, except in February 2001, when the ratio falls to an unprecedented value of 1.77). In the graph 3, it is shown that the change between bull and bear markets results, essentially, from the decreasing of the PGR, while the PLR exhibits more stables values for the entire period. 22 Graph 2 – Evolution of PGR/PLR by Months This graph reports the ratio Proportion of Gains Realized (PGR)/Proportion of Losses Realized (PLR) by month from 1-1-1999 to 31-122002. The number of Realized Gains (RG), Realized Losses (RL), Potential Gains (PG) and Potential Losses (PL) are aggregated across accounts and days. 4,500 4,000 3,500 3,000 2,500 2,000 1,500 1,000 0,500 Jan-99 Feb-99 Mar-99 Apr-99 May-99 Jun-99 Jul-99 Aug-99 Sep-99 Oct-99 Nov-99 Dec-99 Jan-00 Feb-00 Mar-00 Apr-00 May-00 Jun-00 Jul-00 Aug-00 Sep-00 Oct-00 Nov-00 Dec-00 Jan-01 Feb-01 Mar-01 Apr-01 May-01 Jun-01 Jul-01 Aug-01 Sep-01 Oct-01 Nov-01 Dec-01 Jan-02 Feb-02 Mar-02 Apr-02 May-02 Jun-02 Jul-02 Aug-02 Sep-02 Oct-02 Nov-02 Dec-02 0,000 Graph 3 – Evolution of PGR, PLR and its difference by Months This graph reports the Proportion of Gains Realized (PGR), Proportion of Losses Realized (PLR) and its difference by month from 1-1-1999 to 31-12-2002. The number of Realized Gains (RG), Realized Losses (RL), Potential Gains (PG) and Potential Losses (PL) are aggregated across accounts and days. PLR 0,45 PGR 0,40 PGR-PLR 0,35 0,30 0,25 0,20 0,15 0,10 0,05 Jan-99 Feb-99 Mar-99 Apr-99 May-99 Jun-99 Jul-99 Aug-99 Sep-99 Oct-99 Nov-99 Dec-99 Jan-00 Feb-00 Mar-00 Apr-00 May-00 Jun-00 Jul-00 Aug-00 Sep-00 Oct-00 Nov-00 Dec-00 Jan-01 Feb-01 Mar-01 Apr-01 May-01 Jun-01 Jul-01 Aug-01 Sep-01 Oct-01 Nov-01 Dec-01 Jan-02 Feb-02 Mar-02 Apr-02 May-02 Jun-02 Jul-02 Aug-02 Sep-02 Oct-02 Nov-02 Dec-02 0,00 These results, we believe, can only be explained by behavioural reasons. Rationally, one would expect the opposite: momentum strategies in bull periods and contrary strategies in bear periods (contrary strategies conduct to disposition effect). The stronger preference for realising gains and holding losers in bull markets is related to the prospect theory value function, mechanisms of mental accounting and loss aversion. In bull periods, the realization of gains is “easier”: gains are “easily” available and investors will have the desire to realize them because, according to the prospect theory, the satisfaction of a new gain, in terms of value function, is decreasing. This explains why PGR is higher in bull periods. On the 23 contrary, the realization of losses implies the assumption of a wrong decision when the market is going up which is, psychologically, very difficult to accept. As Gervais and Odean (2001) refer, investors learn to be overconfident in bull markets. Then, with reinforced overconfidence investors want to feel the pride of gains (selling winners) and avoid the regret of losses more painful in bull markets (holding their losses and hoping for recovery). Then, investors prefer to hold the stock and avoid the regret of a bad decision. In bear periods, investors are still averse to losses. However, it is easier to accept losses because they can input them to external factors. We found a smooth increase of PLR in the bear period, by comparison to the bull, (from 0.095 to 0.116). When market is decreasing sharply, as it was the case, it becomes easier to accept a loss because investors show a decreasing sensitivity to additional losses and also because the break-even appears very difficult to achieve. Even tough, the increase in PLR is small. As the market is falling down, potential losses increased considerably and, in absolute terms, realised losses increased significantly as well. Simultaneously, the PGR decreases from 0.329 in the bull period to 0.305 in the bear period. When the market is falling, winning stocks are seen as good choices which reinforce the investors’ confidence to keep their decision unchanged. However, according to mental accounting, unrealized gains aren’t considered real gains and, as a result, investors will wait to realise them. Therefore disposition effect still holds. Concluding, we found disposition effect in the bull and bear periods, even tough it is significantly stronger in the bull period. 5.3. The disposition effect and the investors’ sophistication In order to study the disposition effect in relation to the investors’ sophistication, we segment them following three criteria: the trade frequency; the trading volume and the account value. Once it is difficult to select a frontier for deciding whether an investor is a sophisticated one in terms of number of trades, trading volume or account value we divided the sample of investors accounts into two groups, using for frontiers the percentiles 50%, 75%, 90% and 95%. The main purpose is to identify if the disposition effect exists in each group and if there are significant differences between these groups. 24 Tables 8, 9 and 10 exhibit the results for each sophistication criteria: the number of trades; the trading volume and the account value. We found disposition effect for each group and we also found significant differences among these groups: those assumed less sophisticated groups of investors shown significantly stronger disposition effect than the more sophisticated ones. Odean (1998) also found that less frequent traders have stronger disposition effect (Odean only tests the division using the percentile 90%), but the differences we found are much stronger. We also found that, with the increase of the frontier percentile, the disposition effect decreases for the frequent traders group. This means that more active investors are less prone to disposition effect. Consistently, when the division percentile considered is higher the less sophisticated group also exhibits lower disposition effect because we are moving investors from the frequent traders group to the less frequent one. Taking into account the trading volume the conclusions are similar: all groups evidence disposition affect and the groups with lower trading volume show significantly stronger disposition effect. Table 8 – Disposition Effect and Investors’ Trade Frequency This table reports the Proportion of Gains Realized (PGR) and the Proportion of Losses Realized (PLR) from 1-1-1999 to 31-12-2002 for each group of investors. The value of Realized Gains (RG), Realized Losses (RL), Potential Gains (PG) and Potential Losses (PL) are aggregated across accounts and days. The t-statistics calculation assumes that GR, LR, PG and PL result from independent decisions. 50 inferior superior RG 1 331 24 560 RL 727 20 902 PG 1 000 54 553 PL 2 836 174 139 PLR 0.204 0.107 PGR 0.571 0.310 PLR-PGR -0.367 -0.203 PGR/PLR 2.798 2.897 σ(PLR-PGR) 0.01227 0.00179 t-statistic -29.895 -113.698 (PLR-PGR)inf –(PLR-PGR)sup -0.164 t-statistic -13.335 Percentil 75 inferior superior 3 951 21 940 2 644 18 985 3 830 51 723 12 106 164 869 0.179 0.103 0.508 0.298 -0.329 -0.195 2.833 2.884 0.00649 0.00183 -50.634 -106.427 -0.134 -20.644 90 inferior superior 8 436 17 455 6 267 15 362 9 395 46 158 33 031 143 944 0,159 0,096 0,473 0,274 -0,314 -0,178 2,967 2,845 0.00417 0.00192 -75.208 -92.809 -0.136 -32.533 95 inferior superior 11 902 13 989 9 123 12 506 16 224 39 329 52 828 124 147 0,147 0,092 0,423 0,262 -0,276 -0,171 2,874 2,867 0.00327 0.00206 -84.324 -82.991 -0.105 -32.107 25 Table 9 – Disposition Effect and Investors’ Trading Volume This table reports the Proportion of Gains Realized (PGR) and the Proportion of Losses Realized (PLR) from 1-1-1999 to 31-12-2002 for each group of investors. The value of Realized Gains (RG), Realized Losses (RL), Potential Gains (PG) and Potential Losses (PL) are aggregated across accounts and days. The t-statistics calculation assumes that GR, LR, PG and PL result from independent decisions. Percentil RG RL PG PL PLR PGR PLR-PGR PGR/PLR σ(PLR-PGR) t-statistic (PLR-PGR)inf –(PLR-PGR)sup t-statistic 50 inferior 2 091 1 193 2 134 5 919 0.168 0.495 -0.327 2.950 0.008877 -36.857 superior 23 800 20 436 53 419 171 056 0.107 0.308 -0.201 2.888 0.001805 -111.614 -0.126 -14.158 75 inferior superior 5 513 20 378 3 774 17 855 6 185 49 368 19 782 157 193 0.160 0.102 0.471 0.292 -0.311 -0.190 2.942 2.864 0.005197 0.001868 -59.850 -101.821 -0.121 -23.260 90 inferior superior 11 321 14 570 8 397 13 232 15 419 40 134 51 628 125 347 0.140 0.095 0.423 0.266 -0.283 -0.171 3.026 2.789 0.003337 0.002048 -84.956 -83.418 -0.113 -33.752 95 inferior superior 15 317 10 574 11 684 9 945 26 780 28 773 85 218 91 757 0.121 0.098 0.364 0.269 -0.243 -0.171 3.018 2.748 0.002568 0.002421 -94.748 -70.608 -0.072 -28.168 The other criterion used to classify investors’ sophistication, and probably the most accurate one, is the selection based on the account (portfolio) value. Investors with high account value are expected to base their decision making processes on more complex criteria and sophisticated models. Consequently, we should expect them to be less influenced by psychological and behavioural factors. The 5% group of investors with higher account value exhibit a difference between PGR and PLR of 0.14 (with a t-statistic of 51), while the 50% group of investor with lower account value evidence a difference between PGR and PLR of 0.308 (with a t-statistic of 48). This means that the intensity of disposition effect for the lower group is more than twice than the intensity of disposition effect shown by the upper group. Even tough, the disposition effect holds for every group of investors. We conclude that all individual investors seem to prefer realizing gains to losses but the more sophisticated ones evidence a significantly lower disposition effect. Brown, Chappel, Rosa and Walter (2002) also test inventors’ sophistication and disposition effect on the basis of the value of investors’ trades and get similar conclusions. 26 Table 10 – Disposition Effect and Investors’ Account Value This table reports the Proportion of Gains Realized (PGR) and the Proportion of Losses Realized (PLR) from 1-1-1999 to 31-12-2001 for each group of investors. The value of Realized Gains (RG), Realized Losses (RL), Potential Gains (PG) and Potential Losses (PL) are aggregated across accounts and days. The t-statistics calculation assumes that GR, LR, PG and PL result from independent decisions. Percentil RG RL PG PL PLR PGR PLR-PGR PGR/PLR σ(PLR-PGR) t-statistic (PLR-PGR)inf –(PLR-PGR)sup t-statistic 50 inferior 4 168 3 485 3 904 13 221 0.209 0.516 -0.308 2.475 0.006389 -48.167 superior 21 723 18 144 51 649 163 754 0.100 0.296 -0.196 2.968 0.001826 -107.515 -0.111 -17.440 75 inferior superior 8 666 17 225 6 932 14 697 10 296 45 257 34 206 142 769 0.169 0.093 0.457 0.276 -0.289 -0.182 2.712 2.954 0.004061 0.001932 -71.043 -94.374 -0.106 -26.143 90 inferior superior 14 859 11 032 12 110 9 519 21 972 33 581 77 495 99 480 0.135 0.087 0.403 0.247 -0.268 -0.160 2.985 2.832 0.002799 0.002214 -95.823 -72.233 -0.108 -38.694 95 inferior superior 19 422 6 469 15 207 6 422 33 883 21 670 112 244 64 731 0.119 0.090 0.364 0.230 -0.245 -0.140 3.054 2.547 0.002274 0.002728 -107.776 -51.174 -0.105 -46.359 One could say that investors tend to sell winners and hold losers because they expect future reverse movements on prices. But, evidence shows that, on average, this expectation is misleading. This preference for realising winners and holding losers is much more associated with past prices than with expectation about future prices. 27 Conclusions This paper addresses the disposition effect, firstly raised by Shefrin and Statman (1985). According to this behavioural finding, investors tend to less (realising) winner stocks faster and while holding loser stocks longer. The behavioural finance helps to explain such findings, basing investors’ decisions on the apperceived value that investors assign to gains and losses. We use a unique database for the Portuguese market composed by 1496 individual investors’ accounts and 159 406 trades. We find a strong preference for investors selling winners and holding losers and this tendency holds whether the basis of measurement is the number of trades, the trading volume or the turnover. This tendency is far stronger than in other markets previously studied: the Proportion of Gains Realized is 20% higher than the Proportion of Losses Realized. We believe that the main reason for this finding relates with the low level of sophistication of the individual investors. The preference for realizing winners and holding losers holds for each month of the year. As opposed to the evidence found in other markets, the so-called fiscal effect has not enough impact to significantly reduce or invert the disposition effect (in spite of the existence of fiscal incentives under the Portuguese jurisdiction). This specific situation may be due to nonhomogeneous legislation applicable to investors in the sample. This non-homogeneity was also true when we compare subperiods within the entire sampling period. Given our results, we conclude that even some investors decide to realize capital losses for fiscal reasons, there are other investors which are not motivated by fiscal aspects that prefer to realize capital gains at the end of the year (e.g.: to evidence improved performance), balancing each other. This means that mental accounting mechanisms tend to consider as gains those already realized. Therefore, the fiscal effect and the disposition effect counterbalance each other and the pattern of realized gains and losses is not significantly modified on December. Furthermore, we also studied differences in the disposition effect considering market trends (bull/bear). One could expect that, in bull periods, there are more opportunities for realising gains. However, the measure of the disposition effect used considers the realisation of gains and losses compared to the opportunities which exists due to the market trend. We found that, in bull markets, the disposition effect is stronger than in bear markets, even though it is 28 significant for both periods. Rationally, one would expect the opposite: momentum strategies in bull periods and contrary strategies in bear periods. We believe that these results may be explained on the basis of behavioural factors. In bull periods the realization of gains is “easier”: gains are “easily” achievable that investor will desire to realize them because, according to the prospect theory, the satisfaction of a new gain, in terms of value function, is decreasing. On the contrary, the realization of losses implies to assume a wrong decision when the market is going up which is, psychologically, very difficult to accept. Then, investors prefer to hold the stock and avoid the regret of a wrong decision. In bear periods, investors are still averse to losses, but it is easier to accept losses because they can input them to external factors. Finally, we also studied the disposition effect in relation to the investors’ sophistication. In order to identify differences, we divided investors into groups stratified by the following sophistication criteria: number of trades; trading volume and account value. We found that less sophisticated investors exhibit stronger disposition effect, although it is significant in every group of individual investors. Although the Portuguese stock market is no more considered as an emergent market, it may be surprising that it still exhibits some characteristics of those markets and, so, investors tend to reveal low sophisticated investment behaviour, particularly individual investors. 29 Bibliography Brown, P., N. Chappel, R. d. S. Rosa e T. Walter (2002). The Reach of the Disposition Effect: Large Sample Evidence across Investor Classes. European Finance Association Conference, Berlin. Case, K. e R. Shiller (1988). "The Behavior of Home Buyers in Boom and Post-Boom Markets." New England Economic Review: 29-46. Chui, P. (2001). "An Experimental Study of Disposition Effect: Evidence from Macau." Journal of Psychology and Financial Markets 2 (4): 216-222. CMVM (2003). Relatório de Síntese Estatística Sobre a Actividade de Recepção de Ordens por Conta de Outrem por Canal de Recepção, Tipo de Investidor e Mercado de Execução das Ordens. Comissão do Mercado de Valores Mobiliários - Estudos e Documentos. Lisboa. Dhar, R. and N. Zhu (2006). "Up Close and Personal: An Individual Level Analysis of the Disposition Effect." Management Science forthcoming. Dyl, E. (1977). "Capital Gains Taxation and the Year-End Stock Market Behavior." Journal of Finance 32 (1): 165-175. Ferris, S., R. Haugen e A. Makhija (1988). "Predicting Contemporary Volume with Historic Volume at Differential Price Levels: Evidence Supporting the Disposition Effect." Journal of Finance 43 (3): 677-697. Genesove, D. and C. Mayer (2001). "Loss Aversion and Seller Behavior: Evidence from the Housing Market." The Quarterly Journal of Economics 116(4): 1233-1260. Gervais, S. and T. Odean (2001). "Learning to Be Overconfident." The Review of Financial Studies 14(1): 1-27. Grinblatt, M. e M. Keloharju (2000). "The Investment Behavior and Performance of Various Investor Types: A Study of Finland's Unique Data Set." Journal of Financial Economics 55 (1): 43-67. Kahneman, D. e A. Tversky (1979). "Prospect Theory: An Analysis of Decision under Risk." Econometrica 47 (2): 263-292. Kaustia, M. (2000). Can Loss Aversion Move the Market? Evidence from IPO Trading Volume. SSRN Working Paper Series. Kim, K. e J. Nofsinger (2002). The Behavior and Performance of Individual Investors in Japan. Review of Financial Studies Conference, Mannheim. Lakonishok, J. e S. Smidt (1986). "Volume for Winners and Losers: Taxation and Other Motives for Stock Trading." Journal of Finance 41: 951-974. Locke, P. e S. Mann (2000). Do Professional Traders Exhibit Loss Realization Aversion? SSRN Working Paper Series. Odean, T. (1998). "Are Investors Reluctant to Realize Their Losses?" Journal of Finance 53 (5): 1775-1798. Oehler, A., K. Heilmann, V. Lager e M. Oberlander (2002). Dying out or Dying Hard? Disposition Investors in Stock Markets. European Finance Association Conference, Berlin. Schlarbaum, G., W. Lewellen e R. Lease (1978). "Realized Returns on Common Stock Investments: The Experience of Individual Investors." Journal of Business 51 (2): 299-325. Shapira, Z. e I. Venezia (2001). "Patterns of Behavior of Professionally Managed and Independent Investors." Journal of Banking and Finance 25 (8): 1573-1587. Shefrin, H. e M. Statman (1985). "The Disposition to Sell Winners Too Early and Ride Losses Too Long: Theory and Evidence." Journal of Finance 40 (3): 777-790. Thaler, R. (1985). "Mental Accounting and Consumer Choice." Marketing Science 4 (3): 199-214. Weber, M. e C. Camerer (1998). "The Disposition Effect in Securities Trading: An Experimental Analysis." Journal of Economic Behavior and Organization 33 (2): 167-184. 30 View publication stats